In preview
What Databricks is shipping but has not settled. Kept on its own page on purpose: a preview is a thing to investigate, not a thing to build a pipeline or an exam answer on.
Open to everyone and supported for production use, though it can still change. On by default, and an admin can turn it off.
Open to most customers but not for production: support runs through engineering, and the shape of it can still move. On by default on Premium, off on Enterprise.
Invite only, through your account team, and not for production. Here so you know it exists, not so you can plan around it.
Preview Radar
All 70 by number
- 3Configurable recovery period for dropped tables
- 4Git CLI commands in Git folders
- 5Lakehouse Replay
- 6Serverless usage policies
- 7ABAC DENY policies
- 8Domains and subdomains
- 9Environments on classic compute
- 10Fine-grained DML privileges
- 11Governance Hub
- 12Governed agentic app building
- 13Governed tag propagation
- 14Metastore-level ABAC policies
- 15Pages
- 16Bundle deployment history
- 17CustomerLake
- 18Lakewatch
- 19Private network gateway
- 20Serverless compute rate limits
- 21Sharing foreign Delta tables
- 22Sharing metric views
- 23Anomaly detection
- 24Automatic type widening in Auto Loader
- 25Lakebase change data feed
- 26Managed Iceberg materialized views
- 27MySQL and PostgreSQL connectors
- 28Real-time mode in Lakeflow pipelines
- 29Zerobus Ingest into default storage
- 30Full-text search indexes on tables
- 31Kafka-compatible APIs for Zerobus Ingest
- 32Lakebase backup schedule API
- 33Lakehouse//RT
- 34Shopify connector
- 35Unit testing for pipelines
- 36Genie ZeroOps
- 37Lakeflow Jobs health score
- 38Genie Code task in jobs
- 39Lakebase Search
- 40Lakeflow Jobs
- 41LTAP Direct Writes
- 52Databricks-managed MCP servers
- 53MLflow deployment jobs
- 54Agent memory pipelines
- 55Agent services in Unity Catalog
- 56ai_enrich
- 57Full-text search indexes in AI Search
- 58Managed agent memory
- 59MLflow Prompt Registry
- 60MLflow review queues
- 61Production monitoring for GenAI apps
- 62Service policies for AI securables
- 63Unified trace table
- 64Unity Catalog Skills
- 65Unity Gateway
- 66Unity Gateway external service policies
- 67Unity Gateway service policies
- 68Unity Gateway Skills API
- 69AI Runtime
- 70Feature views
Written up here the filled blips above
What the labels actually promise
| Stage | Who can use it | Production use | Support | On by default |
|---|---|---|---|---|
| Private | Invite only | No | Engineering only | No |
| Beta | Most customers | No | Engineering only | Premium yes, Enterprise no |
| Preview | Everyone | Yes | The support team | Yes, with an opt-out |
| GA | Everyone | Yes | The support team | Yes |
The line that matters is between Beta and Public Preview. Public Preview is supported and may be used in production. Beta is not, whatever it looks like in the UI. Databricks publishes these definitions on itsrelease types page.
Every entry, with the detail
Labels come from the documentation page of each feature and are re-checked by the agent. A feature that reaches general availability leaves the radar above and is listed below, with its page kept as it was.